An experiment was conducted to quantify the changes in energy partitioning resulting from grain supplementation in herbage-fed dairy cows at 4 stages during a 670-d lactation. The experiment used 16 lactating Holstein-Friesian cows, with a control and a grain treatment being randomly allocated to 8 cows each. During 4 measurement periods (each of 4d in a metabolism stall and 3d in an indirect calorimeter) beginning at approximately 110, 270, 450, and 560 d in milk (DIM), the energy balance of each cow was measured. Cows in both groups were individually offered freshly cut ryegrass pasture (Lolium hybridum L.) in periods 1 and 3 and ryegrass pasture silage and alfalfa (Medicago sativa L.) hay in periods 2 and 4. In all periods, cows in the grain group were offered an additional 4.4 to 5.0 kg of dry matter of cereal grain/cow per day. Adding grain to the diet increased yields of fat and protein and tended to increase yields of milk and lactose, but did not affect milk composition. Gross energy intake (GEI) declined as lactation progressed. Adding grain to the diet decreased the percentage of GEI in feces and urine, but the extent of these reductions did not change as lactation progressed. Adding grain to the diet similarly reduced the percentage of GEI lost to heat, but again the extent of the reduction remained similar as lactation progressed. The magnitude of the increase in milk energy resulting from grain supplementation did not change with advancing lactation, but tissue energy retention was greater in the first 300 DIM compared with after 300 DIM. For herbage-based diets, CH(4) emissions ranged from 6.2 to 7.6% of GEI, which corresponds to 24.0 to 25.8 g of CH(4)/kg of dry matter intake. For diets supplemented with cereal grains, CH(4) emissions ranged from 6.3 to 7.3% of GEI, which corresponds to 21.6 to 25.2 g of CH(4)/kg of dry matter intake. It was concluded that, for cows producing <24 kg of milk/d and consuming herbage-based diets supplemented with grain, the efficiency of utilizing the additional energy in the grain, as measured by the loss of energy in heat, and its partitioning to milk, did not change as lactation progressed from 110 to 560 DIM.